Comparison of character strings, which can contain mixed single byte and multibyte character data, is handled as follows:
- If expression_1 and expression_2 have different server character sets, then they are converted to the same type. For details, see "Implicit Character-to-Character Translation" in Teradata Vantage™ - Data Types and Literals, B035-1143.
- If expression_1 and expression_2 are of different lengths, the shorter string is padded with enough pad characters to make both the same length.
- Session mode is identified:
In this mode … The default case specification for a character string is … ANSI CASESPECIFIC. Teradata NOT CASESPECIFIC. Unless the CASESPECIFIC phrase is applied to one or both of the expressions, any simple Latin letters in both expression_1 and expression_2 are converted to uppercase before comparison begins.
To override the default case specification of a character expression, apply the CASESPECIFIC or NOT CASESPECIFIC phrase.
- Case specification is determined:
IF … THEN the comparison is … either argument is CASESPECIFIC case specific. both arguments are NOT CASESPECIFIC case blind. - Trailing pad characters are ignored.